The present invention relates to a memory access control method which to divides the data to be stored into a plurality of data fields by a cross-cut method, permitting every data field to be augmented with a respective error correcting code, so that incorrect data can be immediately checked for and corrected when reading. The method greatly improves data access reliability, and saves much data access time and memory space. The invention relates also to a hardware arrangement for the memory access control method.
It is well known that an error may occur when storing data or updating binary data, thereby causing an error message to be read out. When an error message is produced, the normal operation of data will be affected. In order to prevent the occurrence of incorrect data, an ECC (error-correcting code) is used. The ECC add an error correcting code (parity data) for every byte (or word) when storing data, permitting the error correcting code to be stored with the data in the memory. When fetching the data, the error correcting code is read out with the data, and the error correcting code is compared with check data of the fetched data. If the error correcting code is confirmed, the data is accepted for output. If the error correcting code is denied, a syndrome will be produced to recognize the incorrect data, enabling the incorrect data to be corrected. This method of using byte (or word) as a unit for augmented by an error correction code occupies much memory capacity, therefore it reduces the efficiency of the memory. Further, it wastes much time in checking every byte (or word), thereby causing the data operation speed to be reduced.